I think it's fair to say that xISBN (here's an application) has been a fairly popular service. We've done quite a bit of work to make the results better, but little to explore what other types of services could extend or compliment it. I recently posted a similar list to Code4Lib, but thought an entry here might reach a slightly different group.
Here are a few ideas that either we've thought of, or have been proposed to us:
- ISBN to author/title or more extensive work-level metadata
- Adding 'distinctive' information to the response to help selection
- xOCLC: take in an OCLC number and return the OCLC numbers in its FRBR work-set
- xLCCN: do the same thing for LCCNs
- xISSN: do the same thing for ISSNs
- SOAP wrappers around everything
- Expose services via OpenURL 1.0
Some of these would take more work than others. the OCLC number and LCCN sound easy. The author/title information probably isn't too hard, as we are doing similar things in FictionFinder.
Deciding what is distinctive for one ISBN versus the others in a work-set sounds a bit harder. ISSNs sound hard, too. Our FRBR algorithm does group some serials together, but FRBRizing serials is at least slightly contentious and we don't claim to have even a proposed solution, although we do see the need.
We're open to suggestions. It would be nice to have some sort of premium service that would help sustain the free services.